-Z- (Z@Gundam.Com)
Wed, 05 Jan 2000 17:47:15 -0800

At 15:52 1/5/2000 -0500, you wrote:
>-Z- wrote:
>> To which I say "Hey! Wake up and smell the Java!"
>Hmm, well I tested your site with and without NJWin, running on Win98 (N.
>Am. ed.). Using Opera (my fav) and IE v.5.00.2614.3500 (too lazy to
>un-install it). Didn't try Netscape (too lazy to install it). Neither of
>them displayed any of the Turn-A, kanas and kanji properly. So more
>coffee all around. (the IE managed the double quote and delta, so one up
>for MS)

Running IE5 on Win98, you should've been able to browse the lot; that's
what I'm using here at home and on my system at work.

Of course, I've also downloaded all of the language support packs posted on
the Windows Update site. Maybe you need those character sets in the mix
for Unicode to parse properly. Whatever the magic combo may be, I've
replicated it on half a dozen systems (two at home, four at various work
sites, two of those running NT 4.0) so far, so it has to do with the way I
habitually set up IE.

>> If your browser doesn't parse therse characters, in which case you'll get
>> either a question mark or an empty box where the character should be, then
>> you need to consider getting another browser.
>Unicode is a good concept, but in terms of compatibility and acceptance,
>it's pretty behind. To deal with non-Latin characters, you need something
>at the OS level, not the browser level (unless you buy into the prediction
>that browser will merge into the OS). NJWin is a good product because
>it's an extension to the OS, not a browser plug-in (meaning it works (to
>some extent) on Notepad, MS Office etc). Unless MS supports Unicode in
>some future OS (Win2k? Win2k+1?) or a shareware extension (like NJWin)
>comes along, it will be uphill all the way for Unicode. [The Apple OS
>uses Unicode right?]

As noted above, I'm using the IE5 with language packs on Win98 and NT and
I've been able to parse the entire range of Unicode.

Since it comes in at a lower level and preempts the display, could NJWin
actually be interfering with the process?

>Newtype Asylum uses Big5 (Chinese and kanas) on some pages, and S-JIS
>(Japanese) on some others, so it's a mess. Mainland China is pretty
>entrenched with GB encoding, Taiwan, HK and the rest of Chinese world is
>sticking with Big5. Recently I read that S-JIS is a MS bastardization of
>JIS, and is loathed by Japanese scholars. But it's overrunning the older
>more robust encoding because of Win9x/J, is that true?

The Japan Industrial Standard (JIS) is a body of standards, like ISO or
ANSI, governing any number of processes. I don't recall the specific
number that relates to language encoding, but it dates back to the earliest
days of word processing. Shift-JIS is a Microsoft encoding technique using
a double-byte character set (DBCS) that was implemented a decade or so ago,
when 16-bit was as good as it got. Instead of encoding the entire
character, you encode a single-byte base character and a shift value that
points to the second byte. While this may sound awkward and difficult,
it's the same technique used in MS-DOS back in the 8-bit days -- you had an
8-bit program segment prefix and offset to load 16-bit values. If it's
overrunning the purer JIS, it's because it's here now and runs on platforms
where JIS doesn't.

>Honestly I haven't paid much attention to Unicode recently. My hunch is,
>as far as East Asia is concerned, it will be far more expensive to convert
>to Unicode than to fix the Y2K bug.

The best thing about Unicode is that it's a superset of all of the myriad
ISO character sets, allowing them to be mixed on a page. Rather than
encode one page as Western European, another as Japanese, and yet another
as Russian, you can encode one page as Unicode and have all three
interspersed within a single paragraph, displayed in the same font.


Gundam Mailing List Archives are available at http://gundam.aeug.org/

This archive was generated by hypermail 2.0b3 on Fri Jan 07 2000 - 00:45:47 JST